SI/LW sometimes gives the impression of being a doomsday cult,
Because it fits the pattern exactly. If you have top astronomers worrying about meteorite hitting earth, that is astronomy. If you have nonastronomers (with very few astronomers) worrying about meteorite hitting earth, that’s doomsday cult. Or at very best, a vague doomsday cult. edit: Just saying, that’s how I classify, works for me. If you have instances (excluding SIAI) where this method of classification fails in damaging way, I am very interested to hear of them, to update my classification method. I might be misclassifying something. I might just go through the list of things that i classified as cults, and classify some items on that list as non-cult, if the classification method fails.
What complete definition of “cult” are you using here so that I can replace every occurrence of the word by its definition and get a better understanding of your paragraph?
That would be helpful to me as many people use this word in different ways and I don’t know precisely how you use it.
Pretty ordinary meaning: Bunch of people trusting extraordinary claims not backed with any evidence or expert consensus, originating from a charismatic leader who is earning living off cultists. Subtype doomsday. Now, I don’t give any plus or minus points for the leader and living off cultists part, but the general lack of expert concern of the issue is a killer. Experts being people with expertise on relevant subject (but no doomsday experts allowed; has to be something practically useful or at least not all about the doomsday itself. Else you start counting theologians as experts). E.g. for AI risk, the relevant experts may be people with CS accomplishments, the folks who made self driving car, the visual object recognition experts, speech recognition, who developed actual working AI of some kind, etc etc.
I think the problem is conflating different aspects of intelligence into one variable. The three major groups of aspects are:
1: thought/engineering/problem-solving/etc; it can work entirely within mathematical model. This we are making steady progress at.
2: real-world volition, especially the will to form most accurate beliefs of the world. This we don’t know how to solve, and don’t even need to automate. We ourselves aren’t even a shining example of 2, but generally don’t care so much about that. 2 is a hard philosophical problem.
3: Morals.
Even strongly superhuman 1 by itself is entirely harmless, even if very general within the problem space of 1. 2 without 1 can’t invent anything. The 3 may follow from strong 1 and 2 assuming that AI assigns non zero chance to being under test in a simulation, and strong 1 providing enormous resources.
So, what is your human level AI?
It seems to me that people with high capacity for 1, i.e. the engineers and scientists, are so dubious about AI risk because it is pretty clear to them, both internally, and from the AI effort, that 1 doesn’t imply 2 and adding 2 won’t strengthen 1. There isn’t some great issue with 1 that 2 would resolve. The 1 works just fine. If for example we invent awesome automatic software development AI, it will be harmless even if superhuman at programming, and will self improve as much as possible without 2. Not just harmless, there’s no reason why 1-agent plus human are together any less powerful than 1-agent with 2-capability.
Eliezer, it looks like, is very concerned with forming accurate beliefs, i.e. 2-type behaviour, but i don’t see him inventing novel solutions as much. Maybe he’s so scared of the AI because he attributes other people’s problem solving to intellect paralleling his, while it’s more orthogonal. Maybe he imagines that very strongly more-2 agent will somehow be innovative and foom, and he sees a lot of room for improving the 2. Or something along those lines. He is a very unusual person; I don’t know how he thinks. The way I think it is very natural for me that the problem solving does not require wanting to actually do anything real first. That also parallels the software effort because ultimately everyone who is capable of working effectively as innovative software developers are very 1-orientated and don’t see 2 as either necessary or desirable. I don’t think 2 would just suddenly appear out of nothing by some emergence or accident.
Even strongly superhuman 1 by itself is entirely harmless, even if very general within the problem space of 1.
Type 1 intelligence is dangerous as soon as you try to use it for anything practical simply because it is powerful. If you ask it “how can we reduce global temperatures” and “causing a nuclear winter” is in its solution space, it may return that. Powerful tools must be wielded precisely.
See, that’s what is so incredibly irritating about dealing with people who lack any domain specific knowledge. You can’t ask it, “how can we reduce global temperatures” in the real world.
You can ask it how to make a model out of data, you can ask it what to do to the model so that such and such function decreases, it may try nuking this model (inside the model), and generate such solution. You got to actually put a lot of effort, like replicating it’s in-model actions in real world in mindless manner, for this nuking to happen in real world. (and you’ll also have the model visualization to examine, by the way)
What if instead of giving the solution “cause nuclear war” it simply returns a seemingly innocuous solution expected to cause nuclear war? I’m assuming that the modelling portion is a black box so you can’t look inside and see why that solution is expected to lead to a reduction in global temperatures.
If the software is using models we can understand and check ourselves then it isn’t nearly so dangerous.
I’m assuming that the modelling portion is a black box so you can’t look inside and see why that solution is expected to lead to a reduction in global temperatures.
Let’s just assume that mister president sits on nuclear launch button by accident, shall we?
It isn’t an amazing novel philosophical insight that type-1 agents ‘love’ to solve problems in the wrong way. It is fact of life apparent even in the simplest automated software of that kind. You, of course, also have some pretty visualization of what is the scenario where the parameter was minimized or maximized.
edit: also the answers could be really funny. How do we solve global warming? Okay, just abduct the prime minister of china! That should cool the planet off.
It isn’t an amazing novel philosophical insight that type-1 agents ‘love’ to solve problems in the wrong way. It is fact of life apparent even in the simplest automated software of that kind.
Of course it isn’t.
Let’s just assume that mister president sits on nuclear launch button by accident, shall we?
There are machine learning techniques like genetic programming that can result in black-box models. As I stated earlier, I’m not sure humans will ever combine black-box problem solving techniques with self-optimization and attempt to use the product to solve practical problems; I just think it is dangerous to do so once the techniques become powerful enough.
Yup, we seem safe for the moment because we simply lack the ability to create anything dangerous.
Actually your scenario already happened… Fukushima reactor failure: they used computer modelling to simulate tsunami, it was 1960s, the computers were science woo, and if computer said so, then it was true.
For more subtle cases though—see, the problem is substitution of ‘intellectually omnipotent omniscient entity’ for AI. If the AI tells to assassinate foreign official, nobody’s going to do that; got to be starting the nuclear war via butterfly effect, and that’s pretty much intractable.
For more subtle cases though—see, the problem is substitution of ‘intellectually omnipotent omniscient entity’ for AI. If the AI tells to assassinate foreign official, nobody’s going to do that; got to be starting the nuclear war via butterfly effect, and that’s pretty much intractable.
I would prefer our only line of defense not be “most stupid solutions are going to look stupid”. It’s harder to recognize stupid solutions in say, medicine (although there we can verify with empirical data).
It is unclear to me that artificial intelligence adds any risk there, though, that isn’t present from natural stupidity.
Right now, look, so many plastics around us, food additives, and other novel substances. Rising cancer rates even after controlling for age. With all the testing, when you have hundred random things a few bad ones will slip through. Or obesity. This (idiotic solutions) is a problem with technological progress in general.
edit: actually, our all natural intelligence is very prone to quite odd solutions. Say, reproductive drive, secondary sex characteristics, yadda yadda, end result, cosmetic implants. Desire to sell more product, end result, overconsumption. Etc etc.
It’s worth discussing an issue as important as cultishness every so often, but as you might expect, this isn’t the first time Less Wrong has discussed the meme of “SIAI agrees on ideas that most people don’t take seriously? They must be a cult!”
ETA: That is, I’m not dismissing your impression, just saying that the last time this was discussed is relevant.
Less Wrong has discussed the meme of “SIAI agrees on ideas that most people don’t take seriously? They must be a cult!”
Awesome, it has discussed this particular ‘meme’, to prevalence of viral transmission of which your words seem to imply it attributes it’s identification as cult. Has it, however, discussed good Bayesian reasoning and understood the impact of a statistical fact that even when there is a genuine risk (if there is such risk), it is incredibly unlikely that the person most worth listening to will be lacking both academic credentials and any evidence of rounded knowledge, and also be an extreme outlier on degree of belief? There’s also the NPD diagnostic criteria to consider. The probabilities multiply here into an incredibly low probability of extreme on many parameters relevant to cult identification, for a non-cult. (For cults, they don’t multiply up because there is common cause.)
edit: to spell out details: So you start with prior maybe 0.1 probability that doomsday salvation group is noncult (and that is massive benefit of the doubt right here), then you look at the founder being such incredibly unlikely combination of traits for a non-cult doomsday caution advocate but such a typical founder for a cult—on multitude of parameters—and then you fuzzily do some knee jerk Bayesian reasoning (which however can be perfectly well replicated using a calculator instead of neuronal signals), and you end up virtually certain it is cult. That’s if you can do Bayes without doing it explicitly on calculator. Now, the reason I am here, is that I did not take a good look until very recently because I did not care if you guys are a cult or not—the cults can be interesting to argue with. And EY is not a bad guy at all, don’t take me wrong, he himself understands that he’s risking making a cult, and trying very hard NOT to make a cult. That’s very redeeming. I do feel bad for the guy, he happened to let one odd belief through, and then voila, a cult that he didn’t want. Or a semi cult, with some people in it for cult reasons and some not so much. He happened not to have formal education, or notable accomplishments that are easily to know are challenging (like being an author of some computer vision library or what ever really). He has some ideas. The cult-follower-type people are dragged towards those ideas like flies to food.
If it starts worrying more than astronomers do, sure. The few is as in percentile, at same level of the worry.
More generally, if the degree of the belief is negatively correlated with achievements in relevant areas of expertise, then the extreme forms of belief are very likely false. (And just in case: comparing to Galileo is cherry picking. For each Galileo there’s a ton of cranks)
Because it fits the pattern exactly. If you have top astronomers worrying about meteorite hitting earth, that is astronomy. If you have nonastronomers (with very few astronomers) worrying about meteorite hitting earth, that’s doomsday cult. Or at very best, a vague doomsday cult. edit: Just saying, that’s how I classify, works for me. If you have instances (excluding SIAI) where this method of classification fails in damaging way, I am very interested to hear of them, to update my classification method. I might be misclassifying something. I might just go through the list of things that i classified as cults, and classify some items on that list as non-cult, if the classification method fails.
What complete definition of “cult” are you using here so that I can replace every occurrence of the word by its definition and get a better understanding of your paragraph?
That would be helpful to me as many people use this word in different ways and I don’t know precisely how you use it.
Pretty ordinary meaning: Bunch of people trusting extraordinary claims not backed with any evidence or expert consensus, originating from a charismatic leader who is earning living off cultists. Subtype doomsday. Now, I don’t give any plus or minus points for the leader and living off cultists part, but the general lack of expert concern of the issue is a killer. Experts being people with expertise on relevant subject (but no doomsday experts allowed; has to be something practically useful or at least not all about the doomsday itself. Else you start counting theologians as experts). E.g. for AI risk, the relevant experts may be people with CS accomplishments, the folks who made self driving car, the visual object recognition experts, speech recognition, who developed actual working AI of some kind, etc etc.
I wonder what’d happen if we’d train a SPR for cult recognition. http://lesswrong.com/lw/3gv/statistical_prediction_rules_outperform_expert/ SPRs don’t care for any unusual redeeming qualities or special circumstances.
Can you list some non-cult most similar to LW/SIAI ?
There are two claims the conjunction of which must be true in order for a doomsday scenario to be likely:
self-improving human-level AI is dangerous enough
humans are likely to create human-level AI
I am unsure of 2 but believe 1. Do you disagree with 1?
I think the problem is conflating different aspects of intelligence into one variable. The three major groups of aspects are:
1: thought/engineering/problem-solving/etc; it can work entirely within mathematical model. This we are making steady progress at.
2: real-world volition, especially the will to form most accurate beliefs of the world. This we don’t know how to solve, and don’t even need to automate. We ourselves aren’t even a shining example of 2, but generally don’t care so much about that. 2 is a hard philosophical problem.
3: Morals.
Even strongly superhuman 1 by itself is entirely harmless, even if very general within the problem space of 1. 2 without 1 can’t invent anything. The 3 may follow from strong 1 and 2 assuming that AI assigns non zero chance to being under test in a simulation, and strong 1 providing enormous resources.
So, what is your human level AI?
It seems to me that people with high capacity for 1, i.e. the engineers and scientists, are so dubious about AI risk because it is pretty clear to them, both internally, and from the AI effort, that 1 doesn’t imply 2 and adding 2 won’t strengthen 1. There isn’t some great issue with 1 that 2 would resolve. The 1 works just fine. If for example we invent awesome automatic software development AI, it will be harmless even if superhuman at programming, and will self improve as much as possible without 2. Not just harmless, there’s no reason why 1-agent plus human are together any less powerful than 1-agent with 2-capability.
Eliezer, it looks like, is very concerned with forming accurate beliefs, i.e. 2-type behaviour, but i don’t see him inventing novel solutions as much. Maybe he’s so scared of the AI because he attributes other people’s problem solving to intellect paralleling his, while it’s more orthogonal. Maybe he imagines that very strongly more-2 agent will somehow be innovative and foom, and he sees a lot of room for improving the 2. Or something along those lines. He is a very unusual person; I don’t know how he thinks. The way I think it is very natural for me that the problem solving does not require wanting to actually do anything real first. That also parallels the software effort because ultimately everyone who is capable of working effectively as innovative software developers are very 1-orientated and don’t see 2 as either necessary or desirable. I don’t think 2 would just suddenly appear out of nothing by some emergence or accident.
Type 1 intelligence is dangerous as soon as you try to use it for anything practical simply because it is powerful. If you ask it “how can we reduce global temperatures” and “causing a nuclear winter” is in its solution space, it may return that. Powerful tools must be wielded precisely.
See, that’s what is so incredibly irritating about dealing with people who lack any domain specific knowledge. You can’t ask it, “how can we reduce global temperatures” in the real world.
You can ask it how to make a model out of data, you can ask it what to do to the model so that such and such function decreases, it may try nuking this model (inside the model), and generate such solution. You got to actually put a lot of effort, like replicating it’s in-model actions in real world in mindless manner, for this nuking to happen in real world. (and you’ll also have the model visualization to examine, by the way)
What if instead of giving the solution “cause nuclear war” it simply returns a seemingly innocuous solution expected to cause nuclear war? I’m assuming that the modelling portion is a black box so you can’t look inside and see why that solution is expected to lead to a reduction in global temperatures.
If the software is using models we can understand and check ourselves then it isn’t nearly so dangerous.
Let’s just assume that mister president sits on nuclear launch button by accident, shall we?
It isn’t an amazing novel philosophical insight that type-1 agents ‘love’ to solve problems in the wrong way. It is fact of life apparent even in the simplest automated software of that kind. You, of course, also have some pretty visualization of what is the scenario where the parameter was minimized or maximized.
edit: also the answers could be really funny. How do we solve global warming? Okay, just abduct the prime minister of china! That should cool the planet off.
Of course it isn’t.
There are machine learning techniques like genetic programming that can result in black-box models. As I stated earlier, I’m not sure humans will ever combine black-box problem solving techniques with self-optimization and attempt to use the product to solve practical problems; I just think it is dangerous to do so once the techniques become powerful enough.
Which are even more prone to outputting crap solutions even without being superintelligent.
Yup, we seem safe for the moment because we simply lack the ability to create anything dangerous.
Sorry you’re being downvoted. It’s not me.
Actually your scenario already happened… Fukushima reactor failure: they used computer modelling to simulate tsunami, it was 1960s, the computers were science woo, and if computer said so, then it was true.
For more subtle cases though—see, the problem is substitution of ‘intellectually omnipotent omniscient entity’ for AI. If the AI tells to assassinate foreign official, nobody’s going to do that; got to be starting the nuclear war via butterfly effect, and that’s pretty much intractable.
I would prefer our only line of defense not be “most stupid solutions are going to look stupid”. It’s harder to recognize stupid solutions in say, medicine (although there we can verify with empirical data).
It is unclear to me that artificial intelligence adds any risk there, though, that isn’t present from natural stupidity.
Right now, look, so many plastics around us, food additives, and other novel substances. Rising cancer rates even after controlling for age. With all the testing, when you have hundred random things a few bad ones will slip through. Or obesity. This (idiotic solutions) is a problem with technological progress in general.
edit: actually, our all natural intelligence is very prone to quite odd solutions. Say, reproductive drive, secondary sex characteristics, yadda yadda, end result, cosmetic implants. Desire to sell more product, end result, overconsumption. Etc etc.
It’s worth discussing an issue as important as cultishness every so often, but as you might expect, this isn’t the first time Less Wrong has discussed the meme of “SIAI agrees on ideas that most people don’t take seriously? They must be a cult!”
ETA: That is, I’m not dismissing your impression, just saying that the last time this was discussed is relevant.
Awesome, it has discussed this particular ‘meme’, to prevalence of viral transmission of which your words seem to imply it attributes it’s identification as cult. Has it, however, discussed good Bayesian reasoning and understood the impact of a statistical fact that even when there is a genuine risk (if there is such risk), it is incredibly unlikely that the person most worth listening to will be lacking both academic credentials and any evidence of rounded knowledge, and also be an extreme outlier on degree of belief? There’s also the NPD diagnostic criteria to consider. The probabilities multiply here into an incredibly low probability of extreme on many parameters relevant to cult identification, for a non-cult. (For cults, they don’t multiply up because there is common cause.)
edit: to spell out details: So you start with prior maybe 0.1 probability that doomsday salvation group is noncult (and that is massive benefit of the doubt right here), then you look at the founder being such incredibly unlikely combination of traits for a non-cult doomsday caution advocate but such a typical founder for a cult—on multitude of parameters—and then you fuzzily do some knee jerk Bayesian reasoning (which however can be perfectly well replicated using a calculator instead of neuronal signals), and you end up virtually certain it is cult. That’s if you can do Bayes without doing it explicitly on calculator. Now, the reason I am here, is that I did not take a good look until very recently because I did not care if you guys are a cult or not—the cults can be interesting to argue with. And EY is not a bad guy at all, don’t take me wrong, he himself understands that he’s risking making a cult, and trying very hard NOT to make a cult. That’s very redeeming. I do feel bad for the guy, he happened to let one odd belief through, and then voila, a cult that he didn’t want. Or a semi cult, with some people in it for cult reasons and some not so much. He happened not to have formal education, or notable accomplishments that are easily to know are challenging (like being an author of some computer vision library or what ever really). He has some ideas. The cult-follower-type people are dragged towards those ideas like flies to food.
Seems more obviously a doomsday non-cult.
So if the US government worries about meteorites hitting earth, it’s a doomsday cult?
If it starts worrying more than astronomers do, sure. The few is as in percentile, at same level of the worry.
More generally, if the degree of the belief is negatively correlated with achievements in relevant areas of expertise, then the extreme forms of belief are very likely false. (And just in case: comparing to Galileo is cherry picking. For each Galileo there’s a ton of cranks)